【摘要】 人工智能技术的快速发展使得学生借助智能工具生成或抄袭代码的现象日益普遍，教师仅凭代码难以准确判断学生的真实水平，讲解视频作业因此成为一种重要的考察方式。为减轻教师的批阅负担，本研究设计并实现了一套视频作业批阅系统。该系统采用语音识别与基于关键词的浅层语义匹配技术，能够自动将讲解视频转换为文本，并通过提取关键词覆盖率等指标辅助教师评估学生对知识的理解深度。同时，系统还集成了作业全流程管理功能。测试结果表明，系统的语音识别正确率普遍不低于97%，能够为教师提供可靠的辅助参考。更多还原

【Abstract】 The rapid advancement of artificial intelligence technology has made it increasingly common for students to generate or plagiarize code with the help of AI tools. Consequently, it has become difficult for instructors to accurately assess students’ true abilities based solely on code submissions, making video-based programming explanations an important form of evaluation. To reduce the burden of grading such assignments, this study designs and implements a video assignment grading system. The system employs speech recognition and keyword-based shallow semantic matching techniques to automatically transcribe explanation videos into text. By extracting metrics such as keyword coverage, the system assists instructors in evaluating students’ depth of understanding. In addition, the system integrates full-process assignment management functions. Test results show that the system achieves a speech recognition accuracy of no less than 97%, providing reliable auxiliary support for instructors.更多还原